J. P. Mallory Queen’s University (Belfast) Twenty-first century clouds over Indo-European homelands This paper presents the respondent’s general comments to some of the papers of the seminar on the “Indo-European Homeland and Migrations: Linguistics, Archeology and DNA” (Moscow, 12 September, 2012). It briefly examines three homeland models (Neolithic Anato- lia, Near Eastern and Pontic-Caspian) in terms of their ability to address the issues of Indo- European phylogeny (the separation of Anatolian from the rest of the Indo-European lan- guages) and the dispersal of agricultural terms across the Indo-European world. Keywords: Indo-European homeland, Indo-European phylogeny, Indo-European agriculture. In 1900 William Thomson, originally from Belfast and better known as Lord Kelvin, delivered a famous lecture titled ‘Nineteenth-Century Clouds over the Dynamical Theory of Heat and Light’ in which he identified two ‘dark clouds’ that hung over the then clear skies of physics: the failure of science to recover evidence for the aether and its inability to explain black-body radiation. I had already thought of employing this to structure my own comments before I re- ceived the abstracts for this symposium. So I was delighted to see that in the very beginning of T. Gamkrelidze and V. Ivanov’s paper that they allude to the Anatolian god Ni-pá-s which they suggest is cognate with Sanskrit nábhas, Greek néphos, etc, i.e., ‘cloud’. We are met here at a symposium to honour the work of Nikolai Merpert whose own research on the Pontic- Caspian steppe contributed so much to discussions of the origins and dispersal of the Indo- European languages (Merpert 1961; 1965; 1974). A number of the papers delivered at this symposium concern the Indo-European homeland problem and the particular role the Pontic- Caspian steppe may have played within our understanding of the dispersal of the Indo- European languages. For this reason I think it is appropriate that, like my far more eminent Belfast predecessor, I deal with what I perceive to be some of the ‘clouds’ that obscure any of the solutions to the Indo-European homeland problem. And to clarify how serious I think the issues are, I emphasize that I have in mind a dark Russian туча and not a white fluffy облако. Like Lord Kelvin I will limit my discussion to two clouds although if I wanted to prolong the metaphor in earnest I could obscure the skies with as much toxic aerial obstruction as we might associate with the atmosphere of Venus. The speakers at this symposium can generally be seen to support one of the following three ‘solutions’ to the Indo-European homeland problem: 1. The Anatolian Neolithic model. This has been most popularized in the works of Colin Renfrew (1987). It sets Indo-European origins to the Anatolian Neolithic and argues that the spread of the Indo-European languages was part and parcel of the spread of agriculture through Europe in a demographic “wave of advance”. Since its original presentation it has been modified a number of times to deal with some of the more serious criticism directed to- wards it. The revised model (Renfrew 1999) still argues for a movement of farming popula- tions from Anatolia into the Aegean and Balkans extending through central Europe along the Danube drainage (the Linearbandkeramik) and also around the western part of the Black Sea where it carried agriculture and Indo-European languages to the steppelands. The northern Journal of Language Relationship • Вопросы языкового родства • 9 (2013) • Pp. 145–154 • © Mallory J. P., 2013 J. P. Mallory and Atlantic peripheries of Europe are regarded not so much areas of migrant colonization as areas of local acculturation to the new economy. The spread of Indo-European languages into Central and Southern Asia was explained originally by way of two alternative models: a Plan A that saw the Neolithic economy spread eastward from Anatolia towards India (thus the In- dus Civilization might be regarded as Indo-European) or Plan B that explained the Indo- Iranians in terms of a much later migration of Bronze Age peoples from the Asiatic steppe- lands southwards into the territory of southern Central Asia and the Indus. Renfrew eventu- ally abandoned Plan A for Plan B, however, a recent and much publicized solution to the homeland problem by Bouckaert et al (2012), and partly supported by Paul Heggarty at this symposium, appears to argue for a variation of Renfrew’s original Plan A, i.e., a homeland set in Anatolia at the beginning of the Neolithic (7th millennium BCE) with essentially symmetri- cal expansions both west into Europe and east into Asia, although these are not necessarily tied to the initial expansion of farming. 2. The Near Eastern model. The major proponents of this model are the linguists Tamaz Gamkrelidze and Vyacheslav Ivanov (1984) who have been provided with detailed archaeo- logical support by Stanislav Grigoriev (1999; 2002). Here the homeland is set south of the Cau- casus, and Indo-European expansions are set somewhat later than presented in the Anatolian Neolithic model (the spread of farming is not a critical element of the Near East model). A distinctive feature of this model is that the ancient European languages (Balto-Slavic, Ger- manic, Celtic, Italic) are all derived from a Bronze Age migration east of the Caspian through Central Asia. This in effect has created the notion of a secondary homeland located north of the Black and Caspian seas. Another possible variation of this model may be seen in Leonid Sverchkov’s (2012) recent book on Tokharian and, more generally, Indo-European origins in Central Asia. 3. The Pontic-Caspian model. This homeland model, developed in a large number of pub- lications by Marija Gimbutas (e.g., 1991, 351–401) and most recently and extensively argued by David Anthony (2007), locates the homeland in the steppe and forest-steppe regions between the Dnieper and the Volga during the period c 4500–3000 BCE. While there are numerous issues raised by all potential solutions, I wish to simply illus- trate two of the problems, one primarily linguistic and the other archaeological, that constitute ‘clouds’ over any of the solutions. Cloud 1: Linguistic Phylogeny One of the primary tests of the validity of any model of Indo-European origins is whether a solution can account for the phylogeny of the Indo-European languages (Mallory 1997a, 103). In general, archaeologists have been given almost a free hand here because of the lack of agreement among linguists as to the precise shape of the Indo-European family tree and how it should be modeled in space and time. While Indo-Iranian may certainly be seen as a valid subgrouping and Balto-Slavic is certainly a concept embraced by the overwhelming majority of linguists, Graeco-Armenian or Italo-Celtic are areas of considerable debate. So also are some of the broader constructs such as Graeco-Indo-Iranian (with or without Armenian). The position of Tokharian with respect to any other language is similarly a major battleground between those who see it as an orphan, peripheral to the rest of Indo-European, and those who wish to associ- ate it with any number of European (Greek to Germanic) branches. But by and large linguists are agreed on the relative position of one branch: Anatolian was the first language to separate, 146 Twenty-first century clouds over Indo-European homelands either within the framework of Proto-Indo-European or as the co-ordinate half of Indo-Hittite. The essential argument as it is normally presented is that Anatolian lacks a considerable number of features that would characterize Brugmanian Proto-Indo-European (aorist, perfect, subjunc- tive, optative, etc.; Fortsom 2004, 155) and, therefore, its links with an earlier continuum must have been severed before Proto-Indo-European (or the rest of the Indo-European languages) de- veloped in common. This can essentially be explained in one of two ways: 1. The ancestors of the Anatolian languages migrated from the homeland of the proto- language before it developed common Indo-European features. In this model, Anatolian would have preserved an archaic structure while the ancestors of the rest of the Indo- European languages still remained together and evolved later stages of Proto-Indo-European. 2. The ancestors of the Indo-European languages migrated from the homeland of the proto-language. Here it is Proto-Indo-European that moves off to innovate while, presumably, Anatolian was left in the homeland to preserve its archaisms. Obviously we could complicate matters further by proposing a homeland from which both the ancestors of Anatolian and (Proto­)Indo-European migrated in different directions but this would hardly be likely and it would have little bearing on the following discussion. If we apply this test to the three homeland models, we can see how each attempts to sat- isfy this requirement. Pontic-Caspian. In this model the linguistic ancestors of Anatolian are seen to depart earli- est from a homeland north of the Black Sea where they pass through the Balkans (Mallory 1989: 241; Anthony 2007: 259) and, by the beginning of the Bronze Age (depending on which archaeological scenario one wishes to invoke) they enter Anatolia to settle and eventually dominate local non-Indo-European populations such as the Hatti. Later, within the Pontic- Caspian homeland, Brugmanian or mature Proto-Indo-European develops. Subsequent mi- grations carry ancestors of most of the European languages into central and northern Europe while ultimately the linguistic ancestors of the Greeks and Indo-Iranians disperse both west and east during the Bronze Age. These later migrations would also include the ancestors of the Phrygians and Armenians, two other language groups that occupied Anatolia but cannot be regarded as ‘Anatolian’ in the linguistic sense.